It’s like Uber, but for tech support! Working on Anywhere Expert was one of my first big projects at Asurion, and what we learned about finding the right types of experts for our platform was very surprising.
The idea for Anywhere Expert was exciting. We could crowdsource our tech support in a model similar to Uber or other gig economy jobs, give users the ability to set their own schedule, and connect customers with the highest-rated experts to solve their problems. We even had Sprint on board as a pilot partner. I joined right as we were standing up this pilot and we had a very fun opportunity space: how do we find the right types of experts for this job?
First, we needed a welcoming front door. The original version of the site was not great. Very dated, very busy. I wanted to do a full redesign, but first decided to deploy some very quick wins, including a new logo (certainly a work in progress!), making the hero background more inviting, new graphics to showcase how everything worked, and a revamp of the copy to make everything more clear.
With these changes deployed, it was time for a full redesign. I gave myself a week for the changes, including writing the HTML and CSS and deploying to production. I revamped the logo (again), and completely rethought the visual language to elevate the messaging, weave in real user testimonials, and highlight the overall platform process so new users would know exactly what to expect.
These changes went through a few rounds of user testing before finally testing in production, and when it was finally released we saw an increase in conversion and quality of experts.
A new testing approach
The MVP version of the test we gave to potential experts was very barebones. It was a simple Google questionairre that required us to manually follow up with each applicant and conduct a 30 minute interview. We were getting a lot of unqualified leads and the whole process was eating up the team’s time.
With some quick a dirty research and brainstorming, we came up with a segmented approach that started with a short training video that explained the basics and weeded out people who just weren’t interested.
I created a Typeform quiz and embedded it into a custom React wrapper that allowed users to pick back up where they left off if they needed to pause. The quiz was two parts: basic questions that made sure they actually watched the training video, and “test chats” where they had to pretend to answer questions from customers.
Our first version rejected experts that didn’t receive a passing score. They received one chance, and that was it. As an experiment, we released a version that allowed experts to retake the quiz if they missed some questions.
During this time we were also rapidly iterating on our email communications, because our team was getting swamped with emails from potential candidates asking where they were in the process. Even though our final versions were dense, they cut out manual communication touchpoints by nearly half by setting clear expectations and next steps.
To our very great surprise, after measuring the success of experts on the platform long-term, we were able to show a statistical link between high-performing experts and experts who failed the quiz the first time, but took it again and passed.
Our hypothesis: experts who succeeded long-term on the platform tended to be people with tenacity, who may not always know the answer, but will take the time to go figure it out.
As part of the work to create a vision for this platform and our users, I lead the team through a series of brainstorms and working sessions to create experience buckets that were the foundation of our expert journey.
We printed out this huge map, taped it to the wall, and let our team add thoughts, questions, and ideas underneath each section for one week. This helped reshape the focus of some of our teams and inform the 6-12 month roadmap for the platform.